PowerPC Netware <shudder>

Michael Meissner meissner@cygnus.com
Wed Oct 18 16:32:00 GMT 2000


On Wed, Oct 18, 2000 at 12:37:05PM -0700, Geoff Keating wrote:
> Stan Shebs <shebs@apple.com> writes:
> 
> > Hey guys, PowerPC Netware is really really dead, right?  I assume so
> > because the config/netware.h that it would need is long gone, so clearly
> > no one has actually compiled that target in quite some time.  cvs history
> > shows that people are still busily editing away on it, so perhaps we
> > could give them a bit of relief and just evaporate the file?
> 
> How odd.
> 
> It seems that the file config/netware.h is local to Red Hat.  It says:
> 
> /* This file is CYGNUS LOCAL.  */
> 
> /* netware.h -- operating system specific defines to be used when 
>    targeting GCC for some generic NetWare 4 system.
>    Copyright (C) 1993, 1994 Free Software Foundation, Inc.
> 
>    Written by David V. Henkel-Wallace (gumby@cygnus.com)
> 
> The only other netware target is i386, and I don't believe the Red Hat
> i386/netware.h will build, as it has:
> 
> #include "i386/sysv4gdb.h"      /* Build on the base i386 SVR4 configuration */
> #include "netware.h"            /* Then add netware-specific goo. */
> 
> and there doesn't seem to be an i386/sysv4gdb.h.  I can fix that easily.
> 
> After looking at the files, I think probably the best thing to do is
> move the cygnus-local stuff out to the public tree, where perhaps some
> day someone will fix them, and nuke the powerpc stuff, since it can't
> even be configured.

Stan Shebs or Jim Wilson might be able to fill in more bits about this.  The
netware bits were obsolete within Cygnus before I started hacking on the
PowerPC port (figure January 1995).  I vaguely recall that Netware was going to
port to the PowerPC in the great rush to port to the chip (and paid Cygnus to
do the port), and like Windows NT and Solaris, it was killed.  Since the
project was canned before the compiler was done, we never released it.  I
suspect it uses the xcoff object file format, and the standard AIX calling
sequence (primarily because the eabi and ELF bits were completed later, roughly
summer of 1995), though it may use the bastard AIX calling sequence that is
still supported via the -mcall-aix option on the eabi compiler, that is still
in use as far as I can tell.

-- 
Michael Meissner, Red Hat, Inc.
PMB 198, 174 Littleton Road #3, Westford, Massachusetts 01886, USA
Work:	  meissner@redhat.com		phone: +1 978-486-9304
Non-work: meissner@spectacle-pond.org	fax:   +1 978-692-4482


More information about the Gcc mailing list